Starting a new Python project using cookies
In this blog post, I talk about how to quickstart a new project in the popular Python programming language. (I use Python 3k and Ubuntu linux distribution)
I use the cookiecutter command line utility, which helps create project structures for many type of languages, including Python, C, C++, CommonLisp, etc. just to name a few.
Creating the project structure from the command line
I first install cookiecutter:
python3 -m pip install --user cookiecutter
# On ubuntu this works too:
# sudo apt install cookiecutter
Now that I have cookiecutter installed, I checkout the template project released by audreyr the original creator of cookiecutter.
git clone git@github.com:audreyr/cookiecutter-pypackage.git
Then I modify it a bit to reflect my wishes for the project I want to
create. Here I only modified the project metadata contained in
cookiecutter-package/cookiecutter.json
cat cookiecutter-package/cookiecutter.json
#{
# "full_name": "Adrien Oyono",
# "email": "adrienoyono@gmail.com",
# "github_username": "aoyono",
# "project_name": "Python Boilerplate",
# "project_slug": "{{ cookiecutter.project_name.lower().replace(' ', '_').replace('-', '_') }}",
# "project_short_description": "Python Boilerplate contains all the boilerplate you need to create a Python package.",
# "pypi_username": "{{ cookiecutter.github_username }}",
# "version": "0.1.0",
# "use_pytest": "n",
# "use_pypi_deployment_with_travis": "y",
# "add_pyup_badge": "n",
# "command_line_interface": ["Click", "No command-line interface"],
# "create_author_file": "y",
# "open_source_license": ["MIT license", "BSD license", "ISC license", "Apache Software License 2.0", "GNU General Public License v3", "Not open source"]
#}
Then I create my new project from this template:
cookiecutter cookiecutter-package/
I accept all the defaults as responses to the questions asked by the cli, as they are set to the values I put earlier in the metadata file.
A new directory named python_boilerplate is now present in the current
directory and I have a ready to hack python package project repository
ls
# cookiecutter-package python_boilerplate
tree python_boilerplate
# python_boilerplate
# ├── AUTHORS.rst
# ├── CONTRIBUTING.rst
# ├── docs
# │ ├── authors.rst
# │ ├── conf.py
# │ ├── contributing.rst
# │ ├── history.rst
# │ ├── index.rst
# │ ├── installation.rst
# │ ├── make.bat
# │ ├── Makefile
# │ ├── readme.rst
# │ └── usage.rst
# ├── HISTORY.rst
# ├── LICENSE
# ├── Makefile
# ├── MANIFEST. in
# ├── python_boilerplate
# │ ├── cli.py
# │ ├── __init__.py
# │ └── python_boilerplate.py
# ├── README.rst
# ├── requirements_dev.txt
# ├── setup.cfg
# ├── setup.py
# ├── tests
# │ ├── __init__.py
# │ └── test_python_boilerplate.py
# └── tox.ini
#
# 3
# directories, 26
# files
Note:
It would be way more easier if I could skip the
checkout->modify-metadata phase above. It is possible to achieve just
that with cookiecutter. The pre-requisite is to layout a project
structure that I upload to a code hosting platform (Github), so that
every time I need to start a new project, say, a Python package, I just
instruct cookiecutter to check it out for me and create the new
project. I would therefore answer the questions one by one to customize
the project's metadata. It is straightforward:
cookiecutter gh:audreyr/cookiecutter-pypackage
# gh stands for github
When using this method, the git repository is checked out in
~/.cookiecutters/ by default so that you can manage the layouts on
which your projects are based.
That's all folks ! Read the documentation to find out more.